Skip to content

ggml backends interface v1 #547

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 23 commits into from
Oct 6, 2023
Merged

ggml backends interface v1 #547

Show file tree
Hide file tree
Changes from 20 commits
Commits
Show all changes
23 commits
Select commit Hold shift + click to select a range
883f0bc
ggml backends interface v1
slaren Oct 2, 2023
c05714f
move get_alignment from buffer to backend
slaren Oct 3, 2023
da82697
ggml-cuda : fix ggml_cuda_op_mul_mat_vec_q
slaren Oct 3, 2023
3cf87a3
gpt-2 : better check for CPU backend when settings n_threads
slaren Oct 3, 2023
319b4bc
.gitignore : add .clangd
slaren Oct 3, 2023
b527d48
merge master
slaren Oct 4, 2023
abf2669
restore tests/CMakeLists.txt
slaren Oct 4, 2023
a1fd06c
ggml-backend : code style suggestions (#551)
ggerganov Oct 5, 2023
c4cd2d7
gpt-2 : add comments about KV allocation
ggerganov Oct 5, 2023
d8b3efc
add ggml_backend_is_cpu
slaren Oct 5, 2023
3dbc43a
add backend check to ggml_backend_cpu_set_n_threads
slaren Oct 5, 2023
b74ffd5
backend cpu: fix buffer alignment
slaren Oct 5, 2023
b4ec978
fix CUDA_ARCHITECTURES for mmq
slaren Oct 5, 2023
25ce18a
ggml-alloc : better handle view initialization
slaren Oct 5, 2023
b42e19c
ggml-cuda : fix padding clearing
slaren Oct 5, 2023
94b0529
ggml-backend : metal (#552)
ggerganov Oct 6, 2023
ce797df
gpt-2 : take advantage of Metal unified memory
ggerganov Oct 6, 2023
e8bc940
gpt-2 : remove TODO + update comment
ggerganov Oct 6, 2023
5ca14ce
ggml-backend : fix ggml_backend_is_xxx() interface
ggerganov Oct 6, 2023
b22916c
gpt-2 : fix build
ggerganov Oct 6, 2023
01710cc
ggml-cuda : cleanup, fix case for src1 not contiguous
slaren Oct 6, 2023
9cb2626
remove commented code
slaren Oct 6, 2023
1ad7c5e
rename ggml_backend_metal_set_n_threads to n_cb
slaren Oct 6, 2023
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
4 changes: 3 additions & 1 deletion .gitignore
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -6,6 +6,7 @@ build-sanitize-thread/
build-cov/
build-ci-debug/
build-ci-release/
build-cublas/
out/
tmp/
models/
Expand All @@ -15,6 +16,7 @@ compile_commands.json
CMakeSettings.json
.vs/
.vscode/
.clangd

.exrc
.cache
Expand All @@ -32,4 +34,4 @@ zig-cache/

*.sw?

__pycache__/
__pycache__/
15 changes: 15 additions & 0 deletions examples/gpt-2/CMakeLists.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -11,3 +11,18 @@ target_link_libraries(${TEST_TARGET} PRIVATE ggml common common-ggml)
set(TEST_TARGET gpt-2-quantize)
add_executable(${TEST_TARGET} quantize.cpp)
target_link_libraries(${TEST_TARGET} PRIVATE ggml common common-ggml)

#
# For GPU offloading

if (GGML_CUBLAS)
add_compile_definitions(GGML_USE_CUBLAS)
endif()

if (GGML_CLBLAST)
add_compile_definitions(GGML_USE_CLBLAST)
endif()

if (GGML_METAL)
add_compile_definitions(GGML_USE_METAL)
endif()
Loading